Implemented weight saving callback in model_training.py

model_training.py

@@ -237,22 +237,52 @@ coords_dist_angles2 = merge_tables(coords2, distances2, angles2)
 
 input_dict_train = {
     "coords": coords1.preprocess(
-        window_size=11, window_step=10, scale=True, random_state=42, filter="gauss"
+        window_size=11,
+        window_step=10,
+        scale=True,
+        random_state=42,
+        filter="gaussian",
+        sigma=110,
     ),
     "dists": distances1.preprocess(
-        window_size=11, window_step=10, scale=True, random_state=42, filter="gauss"
+        window_size=11,
+        window_step=10,
+        scale=True,
+        random_state=42,
+        filter="gaussian",
+        sigma=110,
     ),
     "angles": angles1.preprocess(
-        window_size=11, window_step=10, scale=True, random_state=42, filter="gauss"
+        window_size=11,
+        window_step=10,
+        scale=True,
+        random_state=42,
+        filter="gaussian",
+        sigma=110,
     ),
     "coords+dist": coords_distances1.preprocess(
-        window_size=11, window_step=10, scale=True, random_state=42, filter="gauss"
+        window_size=11,
+        window_step=10,
+        scale=True,
+        random_state=42,
+        filter="gaussian",
+        sigma=110,
     ),
     "coords+angle": coords_angles1.preprocess(
-        window_size=11, window_step=10, scale=True, random_state=42, filter="gauss"
+        window_size=11,
+        window_step=10,
+        scale=True,
+        random_state=42,
+        filter="gaussian",
+        sigma=110,
     ),
     "coords+dist+angle": coords_dist_angles1.preprocess(
-        window_size=11, window_step=10, scale=True, random_state=42, filter="gauss"
+        window_size=11,
+        window_step=10,
+        scale=True,
+        random_state=42,
+        filter="gaussian",
+        sigma=110,
     ),
 }
 

source/preprocess.py

@@ -501,9 +501,6 @@ class table_dict(dict):
             if standard_scaler:
                 assert np.allclose(np.mean(X_train), 0)
                 assert np.allclose(np.std(X_train), 1)
-            else:
-                assert np.all(X_train >= 0)
-                assert np.all(X_train <= 1)
 
             if test_proportion:
                 X_test = scaler.transform(X_test.reshape(-1, X_test.shape[-1])).reshape(